| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
|
|
If `MRB_INT16` is specified, the variable length argument `mrb_int` is
converted to `int`.
|
|
|
|
The binary sizes (gems are only `mruby-bin-mruby`) are reduced slightly in
my environment than before the introduction of new specifiers/modifiers
(5116789a) with this change.
------------+-------------------+-------------------+--------
BINARY | BEFORE (5116789a) | AFTER (This PR) | RATIO
------------+-------------------+-------------------+--------
mruby | 593416 bytes | 593208 bytes | -0.04%
libmruby.a | 769048 bytes | 767264 bytes | -0.23%
------------+-------------------+-------------------+--------
BTW, I accidentally changed `tasks/toolchains/visualcpp.rake` at #4613,
so I put it back.
|
|
- `size_t` is more commonly used.
- `len` argument of `mrb_str_new()` is `size_t`.
NOTE:
The test for `%l` is temporarily disabled because adding a new type to
`mrbgems/mruby-test/vformat.c` causes an error (memory error?) on Visual
Studio 2017 in AppVeyor.
|
|
It potentially not work when `mrb_float` is `float` because `float` variable
in variable length arguments is promoted to `double`.
Also I fixed build with `MRB_WITHOUT_FLOAT`.
|
|
It potentially breaks, for example, in the case of `mrb_int` is 64-bit
and more smaller type is passed by `%d`. In fact, the problem could
become apparent when I used `%d` to `backtrace_location::lineno` in
`src/backtrace.c:mrb_unpack_backtrace()` on AppVeyor.
Therefore, change `%d` for `int` (not `mrb_int`) so that it can be
used mostly without casting.
|
|
Format sequence syntax:
%[modifier]specifier
Modifiers:
----------+------------------------------------------------------------
Modifier | Meaning
----------+------------------------------------------------------------
! | Convert to string by corresponding `inspect` instead of
| corresponding `to_s`.
----------+------------------------------------------------------------
Specifiers:
----------+----------------+--------------------------------------------
Specifier | Argument Type | Note
----------+----------------+--------------------------------------------
c | char |
d,i | mrb_int |
f | mrb_float |
l | char*, mrb_int | Arguments are string and length.
n | mrb_sym |
s | char* | Argument is NUL terminated string.
t | mrb_value | Convert to type (class) of object.
v,S | mrb_value |
C | struct RClass* |
T | mrb_value | Convert to real type (class) of object.
Y | mrb_value | Same as `!v` if argument is `true`, `false`
| | or `nil`, otherwise same as `T`.
% | - | Convert to percent sign itself (no argument
| | taken).
----------+----------------+--------------------------------------------
This change will increase the binary size, but replacing all format strings
with new specifiers/modifiers will decrease the size because it reduces
inline expansion of `mrb_obj_value()`, etc. at the caller.
|
|
|
|
The addresses for packed inline symbols reference `mrb->symbuf` that
could be overridden by the later call of `mrb_sym2name_len`. Since
file names in call stack information are kept as symbols, keeping the
address in the C structures could cause problems like #4342.
This changes small incompatible changes in function prototypes:
* `mrb_parser_get_filename`: return value changed to `mrb_sym`.
* `mrb_debug_get_filename`: add `mrb_state*` as a first argument.
* `mrb_debug_get_line`: ditto.
I believe above functions are almost internal, and no third-party
mrbgem use them.
|
|
|
|
|
|
`exc_debug_info()` and `mrb_keep_backtrace()` raise `FrozenError`
if the exception is frozen and lead to infinite loop.
|
|
|
|
|
|
|
|
|
|
to 'int', possible loss of data
|
|
'int', possible loss of data
|
|
|
|
|
|
|
|
For readability's sake.
|
|
And arena-overflow error as well. They should not be caught by
`rescue` by default.
|
|
|
|
Instead of preserving a backtrace in `mrb_state`, `mrb_exc_set`
keeps packed backtrace in an exception object. `#backtrace` unpacks
it to an array of strings.
|
|
|
|
|
|
ref #3492 #3515 #3517
|
|
Otherwise line number information is lacked from exceptions
raised in VM, e.g. "super called outside of method".
|
|
|
|
|
|
|
|
|
|
|
|
The change may reduce flexibility, but I believe no one wants
that level of flexibility here.
|
|
|
|
This change reduce flexibility but makes mruby simpler and faster.
|
|
|
|
|
|
Fix broken MRB_INT64
|
|
|
|
|
|
|
|
|
|
PR #3293 just checks for NoMethodError.
|
|
GitHub: fix #3122
It reverts #3126. #3126 fixes the segmentation fault but generates
broken backtrace.
This change fixes the segmentation fault and generates correct
backtrace. The strategy of this change is "generating backtrace while
irep is alive".
/tmp/test.rb:
def gen
e0 = nil
begin
1.times {
raise 'foobar'
}
rescue => e
e0 = e
end
e0
end
e = gen
GC.start
gen
GC.start
puts e.backtrace.join("\n")
Run:
% bin/mruby /tmp/test.rb
/tmp/test.rb:5:in Object.gen
/home/kou/work/ruby/mruby.kou/mrblib/numeric.rb:77:in Integral#times
/tmp/test.rb:4:in Object.gen
/tmp/test.rb:13
FYI:
% ruby -v /tmp/test.rb
ruby 2.3.0p0 (2015-12-25) [x86_64-linux-gnu]
/tmp/test.rb:5:in `block in gen'
/tmp/test.rb:4:in `times'
/tmp/test.rb:4:in `gen'
/tmp/test.rb:13:in `<main>'
|
|
GitHub: fix #2902, #2917
The current implementation traverses stack to retrieve backtrace. But
stack will be changed when some operations are occurred. It means that
backtrace may be broken after some operations.
This change (1) saves the minimum information to retrieve backtrace when
exception is raised and (2) restores backtrace from the minimum
information when backtrace is needed. It reduces overhead for creating
backtrace Ruby objects.
The space for the minimum information is reused by multiple
exceptions. So memory allocation isn't occurred for each exception.
|
